Electronics Spectra - SMS Lucknow

SMS Institute of Technology, L ucknow
Department of Electronics & Co mmunication
of modular Robots.
SCALING AND DESIGN
PRINCIPLES
A fundamental requirement of
Claytronics is that the system must
scale to very large numbers of interacting
catoms. We have the fol lowing
four design principles:
1. Each catom should be self-contained,
in the sense of posses s-
ing everything necessary for performing
its own computation,
communication, sensing, actuation,
locomotion, and adhesion.
2. To support efficient routing o f
power and avoid excessive heat
dissipation, no static power
3. The coordination of the catoms
should be performed via local control.
In particular, no computation
external to the ensemble should
be necessary for individual catom
execution.
4. For economic viability,
manufacturability, and reliability,
catoms should contain no moving
parts.
HARDWARE
At the macro-scale, catoms
have a diameter > 1cm and weigh
many tens of grams. In light o f the
design principles stated above , the
only viable force that can be used to
move and adhere catoms is magnetic;
which sets a lower limit on th e size
and weight of a catom as the m agnets
have considerable weight and volume.
Furthermore, the circuitr y
needed for the high currents n ecessary
to switch the magnets increases
the weight. At this scale, it may not
be possible to adhere to the “no static
power” design principle. Our current
prototype, as shown in Figure 2, is a
system composed of catoms that only
operate in two dimensions. In this
case gravity holds the individual catoms
to a surface and we do not hav e to
deal with the adhesion problem.
Much of the weight and size in a
macro-scale catom comes from packaging,
e.g., chip packages, pins, wires,
PCBs, etc..
In fact, in our planer prototy pe
currently under construction we estimate
that more than 20% of the
weight is packaging and more t han
77% goes to the magnets and th eir
support circuits. This is partly because
we are hand-assembling the catoms.
Our next version will use machine assembly
which will allow us to shrink
the packaging, resulting in an estimated
savings of twofold in we ight
and fivefold in volume.
Micro-scale catoms have diameters
between 1mm and 1cm and weightless
than 1 gram. Almost all packaging
is eliminated and the catom is co n-
structed by bonding VLSI dies directly
to MEMS-based sensor and actuation
dies.
THE SIDE AND TOP
VIEWS OF A
PARTIALLY
ASSEMBLED PLANAR
PROTOTYPE CATOM
The forces necessary to move a
catom are now sufficiently small that
electrostatic forces become an option.
Fig. 2
Macro-scale catoms would requi re
electric field strengths in excess of the
dielectric breakdown of air. Programmable
nanofiber adhesives (PNA), can
be combined with electrostatic actuation
to attach catoms without using
any static power. PNA is the next step
in bio-mimetic adhesives which mainly
use vander waals effects similar to that
of the Gecko. The active components
for micro-scale catoms are within the
realm of current engineering practices
using micro and nanoscale fabrication
techniques. The container coul d be
manufactured by casting a sphe rical
transparent polymer film and sputtering
indium-tin-oxide (ITO) ele ctrode
patches on the film. The patch es of
ITO would be used both to conn ect
catoms for distributing power as well
as used to create capacitive coupling
between catoms in order to con trol
their movement.
The next discontinuity occurs
when catoms are manufactured u s-
ing nanotechnology, e.g., as in chemically
assembled electronic nano-technology.
In this regime, the catoms are
small enough, e.g., < 10 microns in
diameter, that they are aerosol. While
this is currently beyond the state-ofthe-art
in manufacturing further scaling
of lithographic features in VLSI ,
and advances in MEMS capabilities combined
with advances in Nano-technology
will enable the integrate d construction
of such catoms.
In addition to capabilities, the different
regimes (macro, micro, and
nano) have significantly different economics.
Macro-scale catoms require the assembly
of multiple parts into a single
unit. We expect that this will make
the realization of life-size s ynthetic
reality prohibitive due to the cost per
catom. The micro-scale catoms may
also require assembly, but with many
fewer parts, e.g., no boards and the
elimination of magnets in favor of electrostatics.
At this scale the entire
catom will be constructed using a parallel
process, e.g., photolithography.
Just as VLSI-based computers are
commonplace (as opposed to vacuum
tube based computers), in this regime,
catoms are inexpensive enough
that synthetic reality, though expensive,
becomes viable. For exam ple,
assume a 2mm catom controlled by
3mm of silicon. If half the cost is in
the silicon, then, to construc t a human
sized form with an average depth
of 10 catoms would require app roximately
1.25x10.7 catoms or less than
one million dollars. In the nano regime
they will become sufficiently inexpensive
that synthetic reality co uld be
used not only for applications such as
telepario, but could be as ubiquitous
as embedded processors are today.
SOFTWARE
The essence of claytronics—a massively
distributed system composed of
numerous resource-limited catoms—
16 Electronics Spectra, 2010